Behringer DCX2496 digital X-over

At this time the system is using the gain control VST from BlueCat audio controlled via MIDI to control the system volume. Seems to work well but I'll setup the system using Nyquist EQ as the gain control and see if I can hear any difference. I spend a lot of time listening to music at low levels so good performance at higher attenuation levels is plus.

Gary

I'm interested if you come to the same findings and conclusions !


Michael
 
Transformers

Replacing the switcher with an analog supply is a pretty standard mod. For those in North America, Michael Mardis does that using special made (& very special) O'Netics transformers.

dave

The transformers are used to couple the analog output directly out of the dac chip, eliminating all of the active analog stages in the DCX. Other people use caps. I don't know of too many people replacing the switch mode power supply. I scoped mine once and found it to be dead silent at the 200uV limit of my cheap scope. Radiated emissions are noticeable with the lid off but seem to be fully contained within the shielding partitions once the lid is reinstalled. SMPS technology is quite mature now and has sonic advantages over linear supplies especially for power amps. I have just switched to MeanWell SMPS for my power amps and will never look back. The SMPS make the amps sound more powerful and much faster.
 
I have just switched to MeanWell SMPS for my power amps and will never look back. The SMPS make the amps sound more powerful and much faster.

Which ones do you use ?
I've looked up MeanWell page and found them to supply a extraordinary wide palette of SMPS - but nothing dedicated for audio ?

Did you need any additional filters at AC side or DC side ?

Michael
 
MeanWell S-350-36 SMPS sounds great powering my amps

Which ones do you use ?
I've looked up MeanWell page and found them to supply a extraordinary wide palette of SMPS - but nothing dedicated for audio ?

Did you need any additional filters at AC side or DC side ?

Michael

I have switched to a modified version of these amps which sound amazing if you are willing to run a less intrusive output filter.
2*100 watt @ 4ohm TK2050 Class-D Audio Amplifier Board_Sure Electronics' Webstore
See the mod thread here.
http://www.diyaudio.com/forums/class-d/143669-sure-electronics-new-tripath-board-tc2000-tp2050.html
I have finally found a great sounding off the shelf, shielded inductor and am doing listening for the final values of the filter as we speak. Look at my photo gallery to see some of the coils I have tried.
These power supplies are a fantastic deal.
36V DC 9.7A 349.2W Regulated Switching Power Supply_Industrial Power Adapter_Boxed Industrial Power Adapter and Chargers and Transformers_Measure Power Adapters_Sure Electronics' Webstore
No mods necessary. I'm using single rail but they should be able to run +- as the outputs can float.
Once I get my amps together I want to try a twin DCX version of this with T-Line subs.
http://www.diyaudio.com/forums/mult...le-dipole-beyma-tpl-150-a-53.html#post2040924
 
Administrator
Joined 2004
Paid Member
Hey Gary, good to see you here! Love the avatar.

A nice digital volume control would be great. At the moment I use the digital input only so use the volume control in my player. It's done in 32 bits then converted to 24 or 16 bits as I choose. So far, it's working great for me. I'm willing to be proved wrong - that there is a better way.

I've been working with an ESS DAC chip that has 24 bit digital volume control. It also seems quite nice. Being an 8 channel chip, it could run the whole DCX. But it's probably easier to use a drop-in AKM chip with volume control.
 
What about those input and output mods, is the analog transformer going to actually increase THD% from the original design?

Not only that, but many of the transformers commonly used for this have less primary impedance than the 1K Ohm minimum output load that the DAC is rated for. Overloading the DAC probably incurs some noticeable changes in the sound that some may find an improvement.
 
I received my DCX yesterday. I have tested it for multiple hours and came to a conclusion that it is very, very transparent. And believe me it is not because I had bad/low resolution system or something like that.

DCX takes power from pre-amp socket of Bada LB-5600 power filter and I also put a ferrite in the end of the power cord. The power cord itself is shielded (tinned copper foil, 82% coverage) 3x1,5mm2 HQ cable.

IC's are from Monoprice, which are low resistance and inductance, shielded quality cables. My XTZ AP100 power amps have XLR inputs as do the EAW CAZ1400 sub amplifiers. Only unbalanced connection is from pre out of XTZ Class A 100 D3 --> DCX. There I use RCA male --> XLR male adapter and double shielded, air insulted low capasitance, resistance and inductance RCA IC.

As speaker cables I use Ecoflex 15, which is 3/5" thich coaxial cable with very good techinical specs. (center connector is like 5awg thick, very low resistance, capasitance and inductance cable).

I have put some money for the system and tried to optimize it in many aspects. This being said, I do not hear any corruption from DCX in the signal path, as odd it might sound.

DCX adds hiss a little, but I can't hear it from the listening position. What comes to sound quality, hiss would be the only thing I would like to improve in DCX. Hiss itself does not degrade the sound (maybe in very low listening levels it would), but it would be nice if DCX would ntt add any hiss since my set has very low noise floor (Bbarely audible with ear at the tweeter).



What comes to DCX's other features, I noticed that only way to mess the sound is to use auto-align:eek:. Does your DCX also give different delay-settings every time with auto-align?

According the auto-align, the woofer of RX6s is 1 metre/3,3 feet away from midrange and tweeter? I know that analog x-over can cause phase shifting, but wouldn't that be too much and not taken into account when designing the speaker? I listened to this setting and of course it sounded different, and at least to my ear it sounded not so good than all drivers with same delay. Lower midramge was not so present, maybe because midrange/woofer and woofer were not in the same phase (RX6s use 2,5way desing). The I ran A-A again, and result differed quite good from the first settings. 3rd run of A-A and again different result. 3 different runs, 3 different settings. Distance difference is like 0,5-1 metre, while the mic and speakers have stood still.

Where is the problem? Used mic is Behringer ECM8000 measurement mic.
 
Last edited:
Gary, I remember Lynn having mentioned elsewhere, that you did a output mod on the DCX with some hi-grade transformers by your own.

What I am specifically interested in is, if you've done some quality measurements of such transformers - simple harmonics would do here.

The reason I ask is that I never got around transformers to *not* add significant 3nd order at low frequencies - what I call the "iron sound" by myself.

Can this be overcome by selecting core material ?

To show an example - below we see those distortion.
It may be remarkable that these are pretty heavy toroidal monsters spec'ed for 100W power amps (Plitron / Amplimo pat2100-cfb).

But even more amazing for me was that they show this behaviour even at veeeery modest 760mV rms !!!


thd_30hz_2100.jpg


thd_80hz_2100.jpg


The levels here are –60 dB / -65dB for 30Hz / 90 Hz respectively.
You know, I don't wanna go into a discussion about perception of these 3rd harmonics – just want to know what possibly is "common"

So, what realistically to expect in this regard form much smaller – signal level - types ?

Michael
 
Last edited:
Auto-align is pretty consistent for me as well.

Legis, does your setup still have passive crossover components between the amps and speakers? You might also double-check, and try swapping, the absolute phase of the bass or treble components.

Yes, Monitor Audios still have passive x-overs, I'm planning to sell them and buy speakers without passive x-overs though. Could this be the actual cause for very inconsistent distance measurements? This would be odd, since my Marantz sr5004 was very consistent every time, when I had RX6s plugged to it. I will take 5 measurement in a row now and will give you specifict data soon --->
 
Inputs:
1&4 = subwoofers (no subsonic filter/passive x-over)
2&5 = RX6 woofers (with passive x-over)
3&6 = RX6 midrange/woofers & tweeters (with passive x-over)

25mm = 1 inch

Short delays ON
Long delays ON
Polarities ON

Equipment: Behringer EMC8000 mic


Measurement 1

Delays:
1&4 = 1120mm (3,26ms)
2&5 = 1010mm (2,94ms)
3&6 = 1900mm (5,53ms)

Polarities:

1&4 = inv
2&5 = inv
3&6 = inv

Measurement 2

Delays:
1&4 = 1806mm (5,34ms)
2&5 = 1000mm (2,91ms)
3&6 = 2030mm (5,91ms)

Polarities:

1&4 = inv
2&5 = norm
3&6 = inv

Measurement 3

Delays:
1&4= 1306mm (3,80ms)
2&5 = 692mm (2,01ms)
3&6 = 1374mm (4,00ms)

Polarities:

1&4 = inv
2&5 = norm
3&6 = inv

Measurement 4


Delays:
1&4 = 1962mm (5,71ms)
2&5 = 956mm (2,78ms)
3&6 = 2034mm (5,92ms)

Polarities:

1&4 = inv
2&5 = inv
3&6 = inv

Measurement 5

Delays:
1&4 = 596mm (1,68ms)
2&5 = 448mm (1.30ms)
3&6 = 1460mm (4,25ms)

Polarities:

1&4 = inv
2&5 = inv
3&6 = inv


Variance: (smallest value - greatest value)

1&4 = 596mm - 1962mm; (difference 1,33 metres = approx. 4,5 feet)
2&5 = 448mm - 1010mm; (difference 0,58 metres =
approx. 1,8 feet)
3&6 = 1374mm - 2034mm; (difference 0,66 metres =
approx. 2,2 feet)


Is this variation within normal variance? Could someone alse do the same and report using the same layout (easier to compare)?

I don't get those inverted polarities since all my my speaker cables are right of course. And there is variance even in polarity... :)
 
Last edited:
I get more consistent values by testing only left or right side speakers at a time (since microphone might not exactly in the middle of the speakers). Still it thinks that all speakers are inverted.

I highly think that the setting it persistently proposes (that outputs 2&5 are 80cm - 1m closer than others) sounds "wrong phased". It sounds almost like when you have othes speaker's + and - leads wired wrong. In my case: auto-align FTW :D
 
I'm interested if you come to the same findings and conclusions !


Michael

I've got the system setup so I can switch between the 2 setups, one with teh BlueCat Gain plugin and the other with Nyquist Eq setup for gain control. First impression is I can't tell much if any difference at -30dB. It's hard trying to compare sound quality at low level. The instinct is to turn up the volume to hear more but that defeats testing SQ at low volume... I need more test time late at night when things are quiet.

Have you tried the BlueCat gain control?

It did not take long to decide I don't like the way the system responds to volume changes with Nyquist EQ in place of the BlueCat gain control. With Nyquist EQ you end up with a range of -60dB (or so) to +12dB. There does not seem to be a way to limit the range. With BlueCat I'm able to set the range of the volume control to the MIDI range. I've settled on 0 MIDI = -50 dB and 127 MIDI = 0dB.
On the music server/home theater PC we are using an Electrone 9006 wireless keyboard (Electrone - Problem Solving Peripherals) that has a track ball built in. For MIDI control we are using the virtual MIDI XY controller program vimidi. To get the MIDI controls into Console the program MIDI Yoke is used.

We have the ~ key assigned to volume control. Keys 1 through 6 are used to control the 6 input signal pairs. Hold down the ~ key and move the mouse left and right with the track ball to control the volume. This works great because the volume control is available regardless of what window has focus. I can use the volume control even when the 42" monitor is off. Only gotcha is you don't want to adjust the volume if a text entry box in focus...

With Blue Cat gain control the usefull range of control is full left mouse for minimum volume to full mouse right for loud. With the Nyquist EQ the extra 12dB of gain causes loud to occur when the mouse is only 25% of the way across the screen cutting down the resolution of the volume control.

I'll update after some good quiet listening time.

Gary
 
It has been a while since I did any work on my DCX. I have been in the mood to do something, but didn't find anything interesting so far until recent. As some of the readers recall many pages back, I described my mods in detail.

http://www.diyaudio.com/forums/multi-way/15943-behringer-dcx2496-digital-x-over-28.html#post1370646

http://www.diyaudio.com/forums/pass-labs/76095-x-bosoz-finished.html#post872582


I did digital and analog mods. Both were very rewarding. I didn't do much with power supply as I do not believe there is much to be gain there unless the whole circuitry on the board is changed, what is not feasible. As someone mentioned a few posts back, there are some benefits to the switching supply which I do agree with. Really what would be improvement is bypass to the chips that are on the board and maybe little bit more sophisticated regulators.

Anyways, I have been using my Lundahls in the output 1:2 and I ether had them straight to amp or with CCS XBOSOZ preamps that were giving 13 dB gain. My preference would be to use 1:1 Lundahls with no preamps after, just straight to the amp, but that is not possible with speakers and amps I am using. So since that is the case I wanted to see if I could make something little bit more sophisticated and up to date and to replace XBOSOZ as preamp.

The goal was not to have any caps in the signal path, and still to use my Lundahl 1674 amorphous core transformers. Next I wanted to use jFets. So I enlisted help of the mighty Zen Mod. We decided to utilize Lundahls to provide voltage gain since they are 1:4 and than to provide balanced B1 buffer in order to provide proper impedance to the DACs and to isolate DACs from the rest of circuitry. In my case I need to use the Lundahls in 1:4 configuration, but this circuitry would work with 1:2 as well as with 1:1 Lundahl configuration, but RC configuration needs to be adjusted - Located on secondary of transformers. With 1:4 and buffer I have 18dB gain when the circuit is balanced . As you could see in the schematic, Choky designed Cap Multipliers and very good regulation for the power supply.

So far I completed this circuit on the test board and I am listening it now. It sounds really good. The transparency is even better than what I had before with XBOSOZ, very delicate an open. I am planning to produce circuit boards with the help of another good friend Cviller! This circuit is pretty much universal, not just for Behringer but for any other voltage DAC output. Transformer provide DC block as well as attenuate any RF garbage as they have very flat response up to 48 - 50 KHz. The choice of transformer and voltage gain is dependent of the minimum impedance that could be connected to the DAC.

Here is the schematic, for your review. Just to add in order to avoid the confusion the bottom power supply is for my digitally controlled relay volume control.

Thank you Zen Mod and Cviller!
 

Attachments

  • DAC, Lundahl, Buffer.pdf
    75.1 KB · Views: 138
Here is the measurement taken from the circuit on the test board, and without cap multipliers and regulators. I used Tektronix Power supply, and I do expect even more improvement once mounted on the board and with complete power supply as shown in the schematic. This is measured in 1:4 ratio on the Lundahl. Any other lower ratio combination such as 1:2 and 1:1 would yield even better results.
 

Attachments

  • DACLundahlBuffer.pdf
    119.8 KB · Views: 105
Here is the measurement taken from the circuit on the test board, and without cap multipliers and regulators. I used Tektronix Power supply, and I do expect even more improvement once mounted on the board and with complete power supply as shown in the schematic. This is measured in 1:4 ratio on the Lundahl. Any other lower ratio combination such as 1:2 and 1:1 would yield even better results.

Is the "Intermodulation Distorion" plot - where the 3rd order harmonics I mentioned (plus a whole bunch of odd order harmonics) can be seen as well - calibrated in dBu or dBV ?

Michael